One sobering lesson from my time working in the life sciences is that everything is insanely complex and once you've isolated everything in your problem space into a small, relatively isolated set, you still end up with an insanely complex space.

Essentially it's fractally complex.

Except it is actually quite easy to measure hearing thresholds in nonhumans as well as human babies via auditory brain responses (ABRs).

You are right, but so is the grandparent. A noise, click, or tone-based ABR is very basic in terms of measuring restoration of natural hearing. There are issues of rich spectrotempral aspects of speech, music, and natural sounds that (ideally) would be accounted for. Along with amplification and gain control, to which Outer Hair Cells are a major contributor in the cochlea.

There are various ways to measure these thing in animals and humans physiologically (ABRs, otoacoustic emissions, invasive and noninvasive electrophysiology), and with careful psychophysical experiments. None are perfect or comprehensive. Each method has its own strengths and weaknesses.

But it is indeed more complex than just testing a basic audibility threshold. The same issues come up with audigorams for hearing aid calibration and hearing in noise.

life is "fractally" complex. Sometimes I feel computing is also fractally complex.

Well, there's more to it than binary to text you. I agree with you, in that the signaling perception is probably easy. However, let's say that that frequency is coupled with several others, making for noise for poor discrimination.

For that kind of thing, we're going to need real people who can communicate in detail.

Yeah, fitting and tuning a hearing aid is quite complicated, and there are intelligent people at both sides of the process.

The brain also needs time to "rewire" once it starts accepting more and better input.

"Rotating your eyeball" as a treatment for macular degeneration is a good example. Your brain needs a week or so to "reorient" even though the physical procedure is done in a couple hours.

I guess you can easily damage a rats/monkeys hearing consistently (may we find peace for our sins) and I guess you can easily strap those animals into an MRI or strap an EEG measuring device onto one.

Seems simpler than training a significant amount of them!

You may enjoy: https://youtu.be/tONF9OSUOSw

“How Loud Can Sound Physically Get?” - amazing breakdown of the subjective nature of loudness.

More simply, play a loud sound in a narrow band and check if there's a startle response.

Experimental design in medical trials on animals is extremely complex.